root-tree.c 12.9 KB
Newer Older
Chris Mason's avatar
Chris Mason committed
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18
/*
 * Copyright (C) 2007 Oracle.  All rights reserved.
 *
 * This program is free software; you can redistribute it and/or
 * modify it under the terms of the GNU General Public
 * License v2 as published by the Free Software Foundation.
 *
 * This program is distributed in the hope that it will be useful,
 * but WITHOUT ANY WARRANTY; without even the implied warranty of
 * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
 * General Public License for more details.
 *
 * You should have received a copy of the GNU General Public
 * License along with this program; if not, write to the
 * Free Software Foundation, Inc., 59 Temple Place - Suite 330,
 * Boston, MA 021110-1307, USA.
 */

Sachin Kamat's avatar
Sachin Kamat committed
19
#include <linux/err.h>
20
#include <linux/uuid.h>
21
#include "ctree.h"
22
#include "transaction.h"
23 24 25
#include "disk-io.h"
#include "print-tree.h"

26 27 28 29 30 31 32
/*
 * Read a root item from the tree. In case we detect a root item smaller then
 * sizeof(root_item), we know it's an old version of the root structure and
 * initialize all new fields to zero. The same happens if we detect mismatching
 * generation numbers as then we know the root was once mounted with an older
 * kernel that was not aware of the root item structure change.
 */
33 34
static void btrfs_read_root_item(struct extent_buffer *eb, int slot,
				struct btrfs_root_item *item)
35 36 37 38 39 40 41 42 43 44 45 46 47
{
	uuid_le uuid;
	int len;
	int need_reset = 0;

	len = btrfs_item_size_nr(eb, slot);
	read_extent_buffer(eb, item, btrfs_item_ptr_offset(eb, slot),
			min_t(int, len, (int)sizeof(*item)));
	if (len < sizeof(*item))
		need_reset = 1;
	if (!need_reset && btrfs_root_generation(item)
		!= btrfs_root_generation_v2(item)) {
		if (btrfs_root_generation_v2(item) != 0) {
48 49
			btrfs_warn(eb->fs_info,
					"mismatching "
50 51 52 53
					"generation and generation_v2 "
					"found in root item. This root "
					"was probably mounted with an "
					"older kernel. Resetting all "
54
					"new fields.");
55 56 57 58 59 60 61 62 63 64 65 66 67
		}
		need_reset = 1;
	}
	if (need_reset) {
		memset(&item->generation_v2, 0,
			sizeof(*item) - offsetof(struct btrfs_root_item,
					generation_v2));

		uuid_le_gen(&uuid);
		memcpy(item->uuid, uuid.b, BTRFS_UUID_SIZE);
	}
}

68
/*
69 70 71 72 73
 * btrfs_find_root - lookup the root by the key.
 * root: the root of the root tree
 * search_key: the key to search
 * path: the path we search
 * root_item: the root item of the tree we look for
74
 * root_key: the root key of the tree we look for
75
 *
76
 * If ->offset of 'search_key' is -1ULL, it means we are not sure the offset
77 78 79 80
 * of the search key, just lookup the root with the highest offset for a
 * given objectid.
 *
 * If we find something return 0, otherwise > 0, < 0 on error.
81
 */
82 83 84
int btrfs_find_root(struct btrfs_root *root, struct btrfs_key *search_key,
		    struct btrfs_path *path, struct btrfs_root_item *root_item,
		    struct btrfs_key *root_key)
85
{
86 87
	struct btrfs_key found_key;
	struct extent_buffer *l;
88 89 90
	int ret;
	int slot;

91
	ret = btrfs_search_slot(NULL, root, search_key, path, 0, 0);
92
	if (ret < 0)
93
		return ret;
94

95 96 97 98 99 100 101 102 103
	if (search_key->offset != -1ULL) {	/* the search key is exact */
		if (ret > 0)
			goto out;
	} else {
		BUG_ON(ret == 0);		/* Logical error */
		if (path->slots[0] == 0)
			goto out;
		path->slots[0]--;
		ret = 0;
104
	}
105

106
	l = path->nodes[0];
107 108
	slot = path->slots[0];

109
	btrfs_item_key_to_cpu(l, &found_key, slot);
110
	if (found_key.objectid != search_key->objectid ||
111
	    found_key.type != BTRFS_ROOT_ITEM_KEY) {
112 113 114
		ret = 1;
		goto out;
	}
115

116 117 118 119
	if (root_item)
		btrfs_read_root_item(l, slot, root_item);
	if (root_key)
		memcpy(root_key, &found_key, sizeof(found_key));
120
out:
121
	btrfs_release_path(path);
122 123 124
	return ret;
}

125 126
void btrfs_set_root_node(struct btrfs_root_item *item,
			 struct extent_buffer *node)
127 128 129 130 131 132
{
	btrfs_set_root_bytenr(item, node->start);
	btrfs_set_root_level(item, btrfs_header_level(node));
	btrfs_set_root_generation(item, btrfs_header_generation(node));
}

133 134 135
/*
 * copy the data in 'item' into the btree
 */
136 137 138
int btrfs_update_root(struct btrfs_trans_handle *trans, struct btrfs_root
		      *root, struct btrfs_key *key, struct btrfs_root_item
		      *item)
139
{
140
	struct btrfs_path *path;
141
	struct extent_buffer *l;
142 143
	int ret;
	int slot;
144
	unsigned long ptr;
145
	u32 old_len;
146

147
	path = btrfs_alloc_path();
148 149 150
	if (!path)
		return -ENOMEM;

151
	ret = btrfs_search_slot(trans, root, key, path, 0, 1);
152 153 154 155
	if (ret < 0) {
		btrfs_abort_transaction(trans, root, ret);
		goto out;
	}
156 157 158

	if (ret != 0) {
		btrfs_print_leaf(root, path->nodes[0]);
159
		btrfs_crit(root->fs_info, "unable to update root key %llu %u %llu",
160
		       key->objectid, key->type, key->offset);
161 162 163
		BUG_ON(1);
	}

164
	l = path->nodes[0];
165
	slot = path->slots[0];
166
	ptr = btrfs_item_ptr_offset(l, slot);
167 168 169 170 171 172 173 174 175 176 177
	old_len = btrfs_item_size_nr(l, slot);

	/*
	 * If this is the first time we update the root item which originated
	 * from an older kernel, we need to enlarge the item size to make room
	 * for the added fields.
	 */
	if (old_len < sizeof(*item)) {
		btrfs_release_path(path);
		ret = btrfs_search_slot(trans, root, key, path,
				-1, 1);
178 179 180 181 182
		if (ret < 0) {
			btrfs_abort_transaction(trans, root, ret);
			goto out;
		}

183
		ret = btrfs_del_item(trans, root, path);
184 185 186 187
		if (ret < 0) {
			btrfs_abort_transaction(trans, root, ret);
			goto out;
		}
188 189 190
		btrfs_release_path(path);
		ret = btrfs_insert_empty_item(trans, root, path,
				key, sizeof(*item));
191 192 193 194
		if (ret < 0) {
			btrfs_abort_transaction(trans, root, ret);
			goto out;
		}
195 196 197 198 199 200 201 202 203 204 205
		l = path->nodes[0];
		slot = path->slots[0];
		ptr = btrfs_item_ptr_offset(l, slot);
	}

	/*
	 * Update generation_v2 so at the next mount we know the new root
	 * fields are valid.
	 */
	btrfs_set_root_generation_v2(item, btrfs_root_generation(item));

206
	write_extent_buffer(l, item, ptr, sizeof(*item));
207
	btrfs_mark_buffer_dirty(path->nodes[0]);
208
out:
209
	btrfs_free_path(path);
210 211 212
	return ret;
}

213 214
int btrfs_insert_root(struct btrfs_trans_handle *trans, struct btrfs_root *root,
		      struct btrfs_key *key, struct btrfs_root_item *item)
215
{
216 217 218 219
	/*
	 * Make sure generation v1 and v2 match. See update_root for details.
	 */
	btrfs_set_root_generation_v2(item, btrfs_root_generation(item));
220
	return btrfs_insert_item(trans, root, key, item, sizeof(*item));
221 222
}

223 224 225 226 227
int btrfs_find_orphan_roots(struct btrfs_root *tree_root)
{
	struct extent_buffer *leaf;
	struct btrfs_path *path;
	struct btrfs_key key;
228 229
	struct btrfs_key root_key;
	struct btrfs_root *root;
230 231
	int err = 0;
	int ret;
232 233 234 235
	bool can_recover = true;

	if (tree_root->fs_info->sb->s_flags & MS_RDONLY)
		can_recover = false;
236 237 238 239 240 241 242 243 244

	path = btrfs_alloc_path();
	if (!path)
		return -ENOMEM;

	key.objectid = BTRFS_ORPHAN_OBJECTID;
	key.type = BTRFS_ORPHAN_ITEM_KEY;
	key.offset = 0;

245 246 247
	root_key.type = BTRFS_ROOT_ITEM_KEY;
	root_key.offset = (u64)-1;

248 249 250 251 252 253 254 255 256 257 258 259 260 261 262 263 264 265
	while (1) {
		ret = btrfs_search_slot(NULL, tree_root, &key, path, 0, 0);
		if (ret < 0) {
			err = ret;
			break;
		}

		leaf = path->nodes[0];
		if (path->slots[0] >= btrfs_header_nritems(leaf)) {
			ret = btrfs_next_leaf(tree_root, path);
			if (ret < 0)
				err = ret;
			if (ret != 0)
				break;
			leaf = path->nodes[0];
		}

		btrfs_item_key_to_cpu(leaf, &key, path->slots[0]);
266
		btrfs_release_path(path);
267 268 269 270 271

		if (key.objectid != BTRFS_ORPHAN_OBJECTID ||
		    key.type != BTRFS_ORPHAN_ITEM_KEY)
			break;

272 273 274
		root_key.objectid = key.offset;
		key.offset++;

275
		root = btrfs_read_fs_root(tree_root, &root_key);
Sachin Kamat's avatar
Sachin Kamat committed
276
		err = PTR_ERR_OR_ZERO(root);
277
		if (err && err != -ENOENT) {
278
			break;
279 280 281 282 283 284 285 286
		} else if (err == -ENOENT) {
			struct btrfs_trans_handle *trans;

			btrfs_release_path(path);

			trans = btrfs_join_transaction(tree_root);
			if (IS_ERR(trans)) {
				err = PTR_ERR(trans);
287
				btrfs_handle_fs_error(tree_root->fs_info, err,
288 289 290 291 292 293 294 295
					    "Failed to start trans to delete "
					    "orphan item");
				break;
			}
			err = btrfs_del_orphan_item(trans, tree_root,
						    root_key.objectid);
			btrfs_end_transaction(trans, tree_root);
			if (err) {
296
				btrfs_handle_fs_error(tree_root->fs_info, err,
297 298 299 300 301
					    "Failed to delete root orphan "
					    "item");
				break;
			}
			continue;
302 303 304 305 306
		}

		err = btrfs_init_fs_root(root);
		if (err) {
			btrfs_free_fs_root(root);
307 308 309
			break;
		}

310
		set_bit(BTRFS_ROOT_ORPHAN_ITEM_INSERTED, &root->state);
311 312

		err = btrfs_insert_fs_root(root->fs_info, root);
313 314 315 316 317 318 319 320 321
		/*
		 * The root might have been inserted already, as before we look
		 * for orphan roots, log replay might have happened, which
		 * triggers a transaction commit and qgroup accounting, which
		 * in turn reads and inserts fs roots while doing backref
		 * walking.
		 */
		if (err == -EEXIST)
			err = 0;
322 323
		if (err) {
			btrfs_free_fs_root(root);
324 325
			break;
		}
326 327 328

		if (btrfs_root_refs(&root->root_item) == 0)
			btrfs_add_dead_root(root);
329 330 331 332 333 334
	}

	btrfs_free_path(path);
	return err;
}

335
/* drop the root item for 'key' from 'root' */
336 337
int btrfs_del_root(struct btrfs_trans_handle *trans, struct btrfs_root *root,
		   struct btrfs_key *key)
338
{
339
	struct btrfs_path *path;
340 341
	int ret;

342
	path = btrfs_alloc_path();
Tsutomu Itoh's avatar
Tsutomu Itoh committed
343 344
	if (!path)
		return -ENOMEM;
345
	ret = btrfs_search_slot(trans, root, key, path, -1, 1);
346 347
	if (ret < 0)
		goto out;
348

349
	BUG_ON(ret != 0);
Chris Mason's avatar
Chris Mason committed
350

351
	ret = btrfs_del_item(trans, root, path);
352
out:
353
	btrfs_free_path(path);
354 355
	return ret;
}
356 357 358

int btrfs_del_root_ref(struct btrfs_trans_handle *trans,
		       struct btrfs_root *tree_root,
359 360 361
		       u64 root_id, u64 ref_id, u64 dirid, u64 *sequence,
		       const char *name, int name_len)

362
{
363 364 365
	struct btrfs_path *path;
	struct btrfs_root_ref *ref;
	struct extent_buffer *leaf;
366
	struct btrfs_key key;
367 368
	unsigned long ptr;
	int err = 0;
369 370 371
	int ret;

	path = btrfs_alloc_path();
372 373
	if (!path)
		return -ENOMEM;
374 375

	key.objectid = root_id;
376
	key.type = BTRFS_ROOT_BACKREF_KEY;
377
	key.offset = ref_id;
378
again:
379
	ret = btrfs_search_slot(trans, tree_root, &key, path, -1, 1);
380 381 382 383 384 385 386 387 388 389 390 391 392
	BUG_ON(ret < 0);
	if (ret == 0) {
		leaf = path->nodes[0];
		ref = btrfs_item_ptr(leaf, path->slots[0],
				     struct btrfs_root_ref);

		WARN_ON(btrfs_root_ref_dirid(leaf, ref) != dirid);
		WARN_ON(btrfs_root_ref_name_len(leaf, ref) != name_len);
		ptr = (unsigned long)(ref + 1);
		WARN_ON(memcmp_extent_buffer(leaf, name, ptr, name_len));
		*sequence = btrfs_root_ref_sequence(leaf, ref);

		ret = btrfs_del_item(trans, tree_root, path);
393 394 395 396
		if (ret) {
			err = ret;
			goto out;
		}
397 398 399 400
	} else
		err = -ENOENT;

	if (key.type == BTRFS_ROOT_BACKREF_KEY) {
401
		btrfs_release_path(path);
402 403 404 405 406
		key.objectid = ref_id;
		key.type = BTRFS_ROOT_REF_KEY;
		key.offset = root_id;
		goto again;
	}
407

408
out:
409
	btrfs_free_path(path);
410
	return err;
411 412 413 414 415 416 417 418 419 420 421 422 423 424
}

/*
 * add a btrfs_root_ref item.  type is either BTRFS_ROOT_REF_KEY
 * or BTRFS_ROOT_BACKREF_KEY.
 *
 * The dirid, sequence, name and name_len refer to the directory entry
 * that is referencing the root.
 *
 * For a forward ref, the root_id is the id of the tree referencing
 * the root and ref_id is the id of the subvol  or snapshot.
 *
 * For a back ref the root_id is the id of the subvol or snapshot and
 * ref_id is the id of the tree referencing it.
425 426
 *
 * Will return 0, -ENOMEM, or anything from the CoW path
427 428 429
 */
int btrfs_add_root_ref(struct btrfs_trans_handle *trans,
		       struct btrfs_root *tree_root,
430
		       u64 root_id, u64 ref_id, u64 dirid, u64 sequence,
431 432 433 434 435 436 437 438 439 440
		       const char *name, int name_len)
{
	struct btrfs_key key;
	int ret;
	struct btrfs_path *path;
	struct btrfs_root_ref *ref;
	struct extent_buffer *leaf;
	unsigned long ptr;

	path = btrfs_alloc_path();
441 442
	if (!path)
		return -ENOMEM;
443 444

	key.objectid = root_id;
445
	key.type = BTRFS_ROOT_BACKREF_KEY;
446
	key.offset = ref_id;
447
again:
448 449
	ret = btrfs_insert_empty_item(trans, tree_root, path, &key,
				      sizeof(*ref) + name_len);
450 451 452 453 454
	if (ret) {
		btrfs_abort_transaction(trans, tree_root, ret);
		btrfs_free_path(path);
		return ret;
	}
455 456 457 458 459 460 461 462 463 464

	leaf = path->nodes[0];
	ref = btrfs_item_ptr(leaf, path->slots[0], struct btrfs_root_ref);
	btrfs_set_root_ref_dirid(leaf, ref, dirid);
	btrfs_set_root_ref_sequence(leaf, ref, sequence);
	btrfs_set_root_ref_name_len(leaf, ref, name_len);
	ptr = (unsigned long)(ref + 1);
	write_extent_buffer(leaf, name, ptr, name_len);
	btrfs_mark_buffer_dirty(leaf);

465
	if (key.type == BTRFS_ROOT_BACKREF_KEY) {
466
		btrfs_release_path(path);
467 468 469 470 471 472
		key.objectid = ref_id;
		key.type = BTRFS_ROOT_REF_KEY;
		key.offset = root_id;
		goto again;
	}

473
	btrfs_free_path(path);
474
	return 0;
475
}
476 477 478 479 480 481 482 483 484

/*
 * Old btrfs forgets to init root_item->flags and root_item->byte_limit
 * for subvolumes. To work around this problem, we steal a bit from
 * root_item->inode_item->flags, and use it to indicate if those fields
 * have been properly initialized.
 */
void btrfs_check_and_init_root_item(struct btrfs_root_item *root_item)
{
485
	u64 inode_flags = btrfs_stack_inode_flags(&root_item->inode);
486 487 488

	if (!(inode_flags & BTRFS_INODE_ROOT_ITEM_INIT)) {
		inode_flags |= BTRFS_INODE_ROOT_ITEM_INIT;
489 490 491
		btrfs_set_stack_inode_flags(&root_item->inode, inode_flags);
		btrfs_set_root_flags(root_item, 0);
		btrfs_set_root_limit(root_item, 0);
492 493
	}
}
494 495 496 497 498

void btrfs_update_root_times(struct btrfs_trans_handle *trans,
			     struct btrfs_root *root)
{
	struct btrfs_root_item *item = &root->root_item;
499
	struct timespec ct = current_fs_time(root->fs_info->sb);
500

501
	spin_lock(&root->root_item_lock);
502 503 504
	btrfs_set_root_ctransid(item, trans->transid);
	btrfs_set_stack_timespec_sec(&item->ctime, ct.tv_sec);
	btrfs_set_stack_timespec_nsec(&item->ctime, ct.tv_nsec);
505
	spin_unlock(&root->root_item_lock);
506
}